home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   alt.out-of-body      I guess everyone needs a self-vacation      7,897 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 6,915 of 7,897   
   David Mitchell to Dick Silk - The Computer Tutor   
   Re: Re:qualitative data mining.   
   26 Oct 05 19:07:44   
   
   From: david@edenroad.demon.co.uk   
      
   On Wed, 26 Oct 2005 11:05:44 -0500, Dick Silk - The Computer Tutor wrote:   
      
   > "David Mitchell"  wrote in message   
   > news:pan.2005.10.24.20.33.35.980595@edenroad.demon.co.uk...   
   >> On Mon, 24 Oct 2005 13:21:51 -0500, Dick Silk - The Computer Tutor wrote:   
   >   
   >> So it's, essentially, keyword based, stripping the grammatical information   
   >> out. That's what I thought.   
   >   
   > Then you thought wrong.  "Keyword based", and "[stripped of] grammatical   
   > information" is definitely NOT how the final results appear.  Yes, the "top   
   > 100 words" may "appear" keyword based, but a kwb'd search usually indicates   
   > that you already *have* a keyword and that you're attempting to find data   
   > that correlates with it.  Yes, that can be done, however, we don't approach   
   > these studies with the intent of finding data relating to pre-set keywords.   
   > Rather, we let the data sift out and show *us* what is most important in the   
   > minds of the respondents.  In other words, there are NO "pre-defined"   
   > keywords upon which to base any of the data.  KWs are a byproduct *of* the   
   > data, just as a kw list could be created in this very reply I'm sending you.   
      
   Whether they're predefined is irrelevant.  It's the technique which   
   counts, and that technique strips out "key" words (ie. those with content,   
   as you've described).   
      
   >   
   > Further, when compiling the washed / sorted / quantified results, the actual   
   > quotes which support those themes are gathered and grouped together.  So,   
   > for instance, if enough fans had reported that seats, benches, or bleachers   
   > were uncomfortable, their exact quotes would all show up grouped together.   
   > (Seating could be a major top 10 theme, for instance, and the clients would   
   > be interested in reading everything that had to do with comments on   
   > seating.)  Thus, all grammatical information *is* retained.   
      
   Retained, for examination by humans.  Hence not processed by the computers.   
      
   Try to keep up.   
      
   --   
   =======================================================================   
   = David    --- If you use Microsoft products, you will, inevitably, get   
   = Mitchell --- viruses, so please don't add me to your address book.   
   =======================================================================   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca